Since the discovery that acquired immunodeficiency syndrome (AIDS) was caused by a retrovirus, now called human immunodeficiency virus (HIV), a number of anti-retroviral agents have been identified in this laboratory and elsewhere. Several of these agents, including azidothymidine (AZT), 2;,3'-dideoxycytidine (ddC), and 2',3'-dideoxyinosine (ddi) which have potent anti-HIV activity in vitro have been shown to induce laboratory and clinical improvements in patients with AIDS or related disorders. However, these agents are not cures, and their long term use is limited by the development of toxicity and/or resistance in a substantial proportion of patients. We are now exploring whether a combination regimen combining anti-HIV agents may provide a better therapeutic index in patients than single agents. In certain other areas, for example, the treatment of childhood leukemia, combination therapy has been far more useful than single agent therapy; in fact, cures are rarely possible in childhood leukemia with single-agent therapy but are not infrequently obtained with combination therapy. In the setting of HIV infection, rationales for combination therapy include the avoidance of drug toxicities, the delayed development of resistance, drug synergy, the suppression of opportunistic infections which may be acting as co-factors, and the reversal of drug toxicities by counteracting agents. We have initiated several combination drug trials including AZT with acyclovir; AZT and ddc; AZT with granulocyte-macrophage-colony stimulating factor (GM-CSF); AZT with acyclovir, ddc and ddI; and AZT and ddI. Preliminary results indicate that these regimens can be tolerated and may result in reduced toxicity compared with single agents.